Large language models (LLMs) have become (or are becoming) the engine of content creation, especially in business communication. Professionals use AI tools to draft reports, generate marketing content, summarise data, respond to customers, or create their own personal materials. With this expansion comes an important question: What kind of language do these tools produce?

By examining ChatGPT and Perplexity, I claim that both display a recurring syntactic habit: the reduced relative clause formed with present participles. These are the endings that sound like:

  • “…, reshaping consumer behaviour.”
  • “…, driving efficiency across industries.”
  • “…, influencing future strategy.”

While seemingly minor, these endings are a linguistic fingerprint of AI-generated content, with implications for both style and perception in business communication.

What are relative and reduced relative clauses?

A relative clause is a dependent clause that is attached to an antecedent by a relative pronoun such as who, which, or that:

  • The company introduced a new platform, which improved customer engagement.

A reduced relative clause is a shortened adjective clause produced by omitting the relative pronoun (who, which, or that) and the linking verb (to be) or other auxiliary verbs. It only leaves that main verb in the present/past participle form:

  • The company introduced a new platform, improving customer engagement.

Case study

To illustrate these patterns in action, I posed the following request to two LLMs (ChatGPT and Perplexity) at the same time on 22.08.2025:

Prompt: Explain the importance of cars in three sentences.

ChatGPT:

“Cars are important because they provide convenient and efficient transportation, allowing people to travel long distances quickly. They support economic growth by enabling trade, commuting, and access to jobs and services. Additionally, cars offer personal freedom and independence, making daily life more flexible and connected.”

Perplexity:

“Cars are important because they provide convenient and efficient transportation, allowing people to travel long distances quickly and independently. They play a key role in economic growth by enabling the movement of goods, services, and workers. Additionally, cars offer flexibility and accessibility, improving quality of life by connecting people to work, education, healthcare, and leisure activities.”

To test whether AI humanisation would alter their style, I ran both texts through WriteHuman, and the result was as follows:

Version 1:

“Vehicles hold significance as they deliver prompt and dependable travel, letting individuals cover considerable distances with ease and autonomy. They underpin economic expansion by facilitating commerce, commuting, and the transport of personnel and products. Beyond logistics, cars enhance everyday flexibility, linking households to workplaces, schools, hospitals, and recreational spots and thereby elevating overall living standards.”

Version 2:

“Vehicles matter because they put the power of fast, personalised travel in people’s hands, letting them cover great distances without delay. Through streamlined movement of commuters, goods, and services, they drive commerce forward and connect the labour market. On the personal level, they grant the freedom to build daily schedules without depending on fixed routes, linking home to jobs, schools, clinics, and parks in a single blink of an ignition.”

As can be clearly seen, ChatGPT and Perplexity follow the same three-part structure: mobility, economic role, and social benefits. In each case, two of the three sentences (the first and third sentences) end with reduced clauses. Also, the pattern of reduced clauses was preserved in the AI-humanised outputs.

Why AI prefers reduced relative clauses

It is challenging to pinpoint the exact cause without examining how these tools function internally, but some conclusions can be drawn. First, AI systems tend toward conciseness, and reduced relative clauses serve that purpose well by packing more information into fewer words. Second, AI models generate text by maximising probability distributions. The reduced clause is a statistically “safe” option because it frequently appears across many types of texts, from journalism to research articles.

Implications for business communication

Generally speaking, sentence variation is essential for keeping audiences engaged; repetition of the same syntactic pattern risks monotony and predictability. AI’s tendency toward compressed, recurring structures can fatigue readers over time, while varied syntax signals a human touch and sustains attention. If businesses adopt the AI-generated text as is, recognisable patterns (such as the reduced clauses under study) may quickly become clichés that undermine distinctiveness. By contrast, conscious stylistic choices help preserve a brand’s unique voice.

For business communicators, the challenge lies in avoiding formulaic repetition while shaping messages that resonate with the target audience. After all, the effectiveness of communication depends not only on what is said, but also on how it is said.